x001ax000ex001bx0007x000bx000cx0016x0001x - PDF document

  


 


 


!"

 




    
 


 


 
#
+
 
 
, 
 
&

#x0004;

  
  (
!*
C
   ANTIBIOTIC UPTAKE INGRAM-NEGATIVE BACTERIA Claudio Muheim Antibiotic uptake inGram-negative bacteria Claudio Muheim ©Claudio Muheim, Stockholm University 2017 ISBN print 978-91-7797-041-5ISBN PDF 978-91-7797-042-2 Printed in Sweden by Universitetsservice US-AB, Stockholm 2017 To the decentralized future. YfgM is an Ancillary Subunit of the SecYEG Translocon in Götzke H, Palombo I, Muheim C, Perrody E, Genevaux P, Kudva R, Müller M, Daley DO. (2014) 289(Identification of Putative Substrates for the Periplasmic Escherichia coli Using Quantitative Götzke H, Muheim C, Maarten AF, Heck AJ, Maddalo G, Daley DO. Mol Cell Proteomics (2015) 14(), 216…226Increasing the Permeability of Muheim C, Götzke H, Eriksson AU, Lindberg S, Lauritsen I, Nørholm MHH, Daley DO. Identification of a Fragment-Based Scaffold that Inhibits the Escherichia coliMuheim C, Bakali A, Engström O, Wieslander Å, Daley DO, Widmalm G. Antibiotics (2016) 5( I performed the antibiotic disc diffusion assays and was involved in the interpretation of the results.I cloned the plasmid constructs and performed the acid-stress assays.I was involved in the design of all experiments. I performed all of the experiments

apart from the high-throughput screen. I contributed significantly in writing the final manuscript.I designed and performed all experiments apart from the lipid binding assay and the -Deacylation of LPS. I contributed significantly in writing the final manuscript. The increasing emergence and spread of antibiotic-resistant bacteria is a serious threat to public health. Of particular concern are Gram-negative Escherichia coli, Acinetobacter baumannii or . Some of these strains are resistant to a large number of antibiotics and thus our treatment options are rapidly declining. In addition to the increasing number of antibiotic-resistant bacteria, a major problem is that many of the antibiotics at our disposal are ineffective against Gram-negative bacteria. This is partly due to the properties of the outer membrane (OM) which prevents efficient uptake. The overarching goal of this thesis was to investigate how the OM of the Gram-negative bacterium could be weakened to improve the activity of In the first two papers of my thesis (paper I + II), I investigated the periplasmic chaperone network which consists of the two parallel pathways SurA and Skp/DegP. This network is essential for the integrity of the OM and strains lacking either SurA or Skp are defective in the assembly of the OM, which results in an increased sensitivity towards vancomycin and other antimicrobials. We identified a novel component of the periplasmic chaperone network, name

ly YfgM, and showed that it operates in the same network as Skp and SurA/DegP. In particular, we demonstrated that deletion of YfgM in strains with either a background further compromised the integrity of the OM, as evidenced by an increased sensitivity towards vancomycin. In the remaining two papers of my thesis (paper III + IV), the goal was to characterize small molecules that permeabilize the OM and thus could be used to improve the activity of antibiotics. Towards this goal, we performed a high-throughput screen and identified an inhibitor of the periplasmic chaperone LolA, namely MAC-13243, and showed that it can be used to permeabilize the OM of E. coli (paper III). We further demonstrated that MAC-13243 can be used to potentiate the activity of antibiotics which are normally ineffective against . In the last paper of my thesis (paper IV), we undertook a more specific approach and wanted to identify an inhibitor against the glycosyltransferase WaaG. This enzyme is involved in the synthesis of LPS and genetic inactivation of WaaG results in a defect in the OM, which leads to an increased sensitivity to various antibiotics. In this paper, we identified a small molecular fragment (compound L1) and showed that it can be used to inhibit the activity of WaaG To summarize, this thesis provides novel insights into how the OM of the Gram-negative bacterium E. coli can be weakened by using small molecules. We believe that the two identified

small molecules represent important first steps towards the design of more potent inhibitors that could be used in clinics to enhance the activity of antibiotics. List of publications .................................................................................... viiiAuthors contribution to the publications ................................................. ixAbstract .......................................................................................................... xContents ...................................................................................................... xiiAbbreviations ............................................................................................. xivIntroduction ................................................................................................. 161.1 Antibiotics .......................................................................................... 161.2 Antibiotic resistance ........................................................................... 181.3 The antibiotic crisis ............................................................................ 201.3.1 The lack of novel antimicrobials ................................................ 211.3.2 The excessive use and misuse of antibiotics ............................... 221.3.3 The lack of infection control and surveillance programs ............ 231.4 The emerging threat of Gram-negative bacteria ......................

........... 241.5 Structural overview of the Gram-negative cell envelope ................... 251.5.1 The outer membrane ................................................................... 261.5.2 The periplasmic space ................................................................. 281.5.3 The inner membrane ................................................................... 281.6 Antibiotic transport across the Gram-negative cell envelope ............. 291.6.1 Porin-mediated antibiotic uptake ................................................ 291.6.2 Diffusion across the LPS ............................................................ 301.6.3 Diffusion across the periplasmic space and IM .......................... 311.6.4 Efflux pumps .............................................................................. 33 1.7 Antibiotic adjuvants ........................................................................... 341.8 Additional targets for antibiotic adjuvants ......................................... 37Summary of the papers .............................................................................. 44Conclusions and future perspectives ......................................................... 57Populärvetenskaplig sammanfattning....................................................... 61Acknowledgements ..................................................................................... 62References ....................................

................................................................ 64 A. baumannii ABC ATP-binding cassette BAM -barrel assembly machinery CL Cardiolipin EPIs Efflux pump inhibitors HEP IM Inner membrane KDO 3-deoxy-K. pneumoniae LPS Lipopolysaccharide MATE Multidrug and toxic compound extrusion MIC Minimal Inhibitory Concentration MDR Multidrug resistance MFS Major facilitator superfamily NPN N-phenyl-1-naphthylamine OM Outer membrane OMP Outer membrane protein OS Oligosaccharide N Phenylalanine-arginine -naphthylamide PE Phosphatidylethanolamine PG Phosphatidylglycerol P. aeruginosa Pseudomonas aeruginosaPMBN Polymyxin B nonapeptide RND Resistance-nodulation-division SMR Small multidrug resistance TPR Tetratricopeptide repeat WT Wild type 16 Microbes, such as bacteria and fungi synthesize antibiotics and secrete them into the environment to inhibit growth of competing microbes. Millions of years of chemical warfare between rivalling species formed an arsenal of antibiotics that we use today to treat bacterial infections (1, 2). Based on their antimicrobial activity, antibiotics can be broadly classified into two groups: bacteriostatic antibiotics arrest the growth of bacteria but do not kill them whereas bactericidal antibiotics induce cell death (Figure 1). However, it is often difficult t

o classify an antibiotic into one of these categories since antibiotics can have both bacteriostatic and bactericidal activity. Various factors such as growth conditions, bacterial density, antibiotic concentration or antibiotic exposure time influence the antimicrobial activity (3, 4). Thus, antibiotics are also categorized into different classes, based on their structure, mechanism or spectrum of activity. One of the most prominent classifications is based on the chemical structure, since antibiotics with structural similarities typically show similar effects regarding their activity and toxicity. Figure 1. Antimicrobial effect of bacteriostatic and bactericidal antibiotics.Based on their antimicrobial effect, antibiotics can either inhibit bacterial growth (green curve) or induce cell death (red curve). Various factors such as growth conditions, bacterial density, antibiotic concentration or antibiotic exposure time influence the antimicrobial activity. Figure adapted from (5). Reprinted with permission. Despite approximately 3,000 known antibiotics, most of their cellular targets have not been characterized yet (6). Interestingly, those antibiotic targets which have been identified can be assigned to a few cellular processes (7). These include cellular processes which are essential for the survival of bacteria such as DNA replication, RNA transcription, protein translation, cell wall synthesis, the cell membrane and a few other metabol

ic pathways (Figure 2). One of the most successful antibiotic classes are the -lactams are a class of broad-spectrum antibiotics which inhibit cell wall synthesis (8). In -lactams prevent crosslinking of the peptidoglycan units by inhibiting peptide bond formation which is catalysed by penicillin-binding -lactam antibiotics have been introduced into the clinics which include the penicillins, the cephalosporins, the cephamycins, the carbapenems and the monobactams. However, despite -lactams and other antibiotics have been 18 challenged by the problem of increasing bacterial resistance. As a consequence, these antibiotics need to be constantly evolved in order to evade bacterial resistance. Figure 2. Antibiotic classes inhibit various cellular Most antibiotic classes inhibit only a few cellular processes. These are often conserved among bacteria and include DNA replication, transcription, translation, cell wall synthesis, the cell membrane and a few other metabolic pathways (not shown). In the traditional sense, antibiotic resistance occurs when a bacterial species acquires the ability to resist the action of an antibiotic. This process is also known as acquired resistance and can either happen by mutation of the bacterial chromosome or by horizontal gene transfer of foreign DNA coding for antibiotic resistance elements. Horizontal gene transfer is particularly problematic since it largely contributes to the spread of antibiotic resistance

elements within and even across bacterial species (9). On the other hand, bacterial species can be naturally resistant to certain antibiotics. These species have the ability to withstand the action of an antibiotic due to its inherent structural or functional properties. These properties are encoded in the bacterial genome and ensure that bacteria are intrinsically resistant to The general mechanisms which confer resistance to antibiotics are depicted in Figure 3. These include (1) utilizing an alternative metabolic pathway which bypasses the inhibited one, (2) modification of the antibiotic target that prevents or reduces binding of the antibiotic, (3) overproduction of the antibiotic target, (4) enzymatic inactivation or modification of the antibiotic, (5) active efflux of the antibiotic by efflux pumps and (6) reduced antibiotic uptake by a decreased permeability of the outer membrane (OM) (10). Figure 3. Antibiotic resistance mechanisms. Bacterial cells have general mechanisms to resist the action of an antibiotic. These include (1) bypassing the inhibited pathway, (2) alteration of the target, (3) amplification of the target, (4) enzymatic inactivation or modification of the antibiotic, (5) increased antibiotic efflux and (6) reduced antibiotic uptake. Figure adapted from (10). Reprinted with permission. 20 The discovery of the -lactam antibiotic penicillin by the Scottish scientist Alexander Fleming in 1928 is considered as a key

event in modern medicine (11). After being introduced into the clinics in the 1940s, penicillin saved thousands of lives. For example, it was used to treat pneumococcal wound infections that were caused by battlefield injuries, and as a consequence, the survival rate of injured soldiers significantly improved (12). The great More than 20 new antibiotic classes were discovered from 1940 to 1962 Figure 4. Timeline of antibiotic discovery. Figure adapted from (14). Reprinted with permission. Notably, most antibiotics which are currently in clinical use were discovered during this period. However, despite ongoing efforts, this productive phase abruptly ended in the 1960s and since then only two new antibiotic classes were clinically approved for human use, namely the oxazolidinones in 2000 and the lipopeptides in 2003 (15). These two classes were discovered decades ago (Figure 4) and illustrate how long it can take until a novel antibiotic class enters into the market. It is estimated that we need a further 20 novel antibiotic classes to support modern medicine during the next 50 years (13). There are several reasons which are believed to be the driving forces of the antibiotic crisis. As previously described, the number of novel antibiotic classes entering into the market has been steadily declining over the last few decades. In addition, the number of multidrug resistant bacteria has been increasing due to an excessive use of antibiotics. F

inally, the global spread of multidrug resistant bacteria has been increasing due to an enhanced global connectivity. In the next few sections, I will discuss each of these points and explain some of the major problems. 1.3.1 The lack of novel antimicrobials The antibiotic pipeline is drying out and the situation is deteriorating (16). After the antibiotic gold rush ended in the 1960s, pharmaceutical companies shifted their focus on analogue development because of a gradually declining number of novel antibiotic classes. The toxicity risks of analogues were considered to be more predictable compared to new antibiotic classes (13). Despite the risk of cross-resistance, analogue development yielded a number of new antibiotics that kept up with the emergence of bacterial resistance until the last two decades (13). Since then, however, many pharmaceutical companies have completely abandoned antibiotic drug discovery and development programs. It is no longer considered to be an attractive investment because of regulatory obstacles, increasingly restricted 22 antibiotic use and uncertainties about emerging resistance (17…19). Additional efforts to develop new platforms for antibiotic drug discovery, such as high-throughput screening against defined targets or rational drug design were largely unsuccessful (6, 20). Although there is increasing evidence that pharmaceutical companies are restarting antibiotic drug discovery programs, it remains u

nclear if the antibiotic pipeline can be filled to its current need (21). 1.3.2 The excessive use and misuse of antibioticsAntibiotics are among the most commonly purchased drugs (22). There is clear evidence that antibiotic resistance is a direct consequence of excessive antibiotic use (23…25). It is estimated that about half of all antibiotics prescribed to patients are not required or inadequately prescribed (26). Antibiotic misuse is especially problematic in countries which have no regulations regarding antibiotic dispensation. In these countries, antibiotics are sold without prescription and dispensed by people lacking medical knowledge (27). Another major problem is the use of antibiotics as growth supplements in livestock (28, 29). In this common practice, a sub-therapeutic concentration of antibiotic is mixed to the feed to improve the feed efficiency. Some of these growth promotors belong to antibiotic classes which are used in human medicine and it has been reported that this can lead to cross-resistance (30). Another and often neglected problem is the environmental contamination with antibiotics by pharmaceutical production facilities in developing countries. Antibiotic waste products end up unfiltered in natural waters, which causes a selection pressure on bacteria that eventually leads to the emergence of antibiotic-resistant bacteria (31). programs Antibiotic resistance is an international problem and the increase in global c

onnectivity is partly responsible for the transmission of multidrug resistant bacteria across borders. Multidrug-resistant bacteria often end up in healthcare facilities where they spread among patients and healthcare professionals (32). In addition, there is often a lack of infection control and surveillance programs in these facilities which further promotes the transmission of drug-resistant bacteria (33). As a consequence, thousands of people die every year from hospital-acquired infections (34). Hence, various antimicrobial stewardship programs have been initiated to address this issue (35…37). These programs are coordinated at a national and international level and address points to limit transmission rates within healthcare facilities and community settings as well as monitoring antibiotic resistance and antibiotic 24 It is evident that we are entering into a period with a lack of novel antibiotics to treat bacterial infections. The discovery of novel antibiotics is not keeping up with the emergence of antibiotic resistant bacteria. Although resistance to antibiotics is a general problem, the increasing number of infections caused by multidrug resistant Gram-negative bacteria is alarming (38, 39). Highly or are among the most critical threats (26, 40). These pathogens are of particular concern in both healthcare facilities and community settings (41). It is no longer uncommon that some of these strains are resistant to nearly all a

vailable antibiotics. Thus, the number of treatment options rapidly decline and clinicians are forced to use older drugs, such as the last-resort antibiotic colistin, which has been associated with neuro- and nephrotoxicity (42). The situation is alarming and we need new strategies to tackle the problem of Gram-negative resistance. It is worth mentioning that the unique architecture of the Gram-negative cell envelope is partly responsible for the problem. A combination of limited antibiotic uptake across the OM and efflux of antibiotics by multidrug efflux pumps limit the activity of many antimicrobials. Since the work covered in this doctoral thesis has aimed to focus on the antibiotic uptake problem, the following sections will introduce a structural overview of the Gram-negative cell envelope whilst describing the properties that make the cell envelope largely impermeable to most antibiotics. 1.5 Structural overview of the Gram-negative cell The cell envelope is a complex structure which is involved in various cellular processes, such as protection of the cell and maintenance of cellular shape (43). It is composed of three essential layers, namely an OM that faces towards the exterior, a periplasmic space that includes a thin layer of peptidoglycan and an inner membrane (IM) that separates the periplasm from the cytosol (Figure 5). Figure 5. The cell envelope of Gram-negative bacteria. The Gram-negative cell envelope consists of an OM

that provides a barrier to the extracellular environment, a periplasmic space which includes a thin layer of peptidoglycan and an IM that separates the periplasmic space from the cytosol. The OM contains lipopolysaccharides (LPS) and various outer membrane proteins (OMPs), which have a characteristic -barrel structure. The IM contains integral membrane proteins as well as peripherally attached membrane proteins. In addition, both membranes contain various lipoproteins which face towards the periplasmic side. Adapted from (44). Reprinted with permission. 26 The OM is the outermost barrier, which separates the extracellular environment from the inner compartment of the bacterial cell. It is an asymmetric lipid bilayer that consists of two leaflets: the inner leaflet predominantly contains phospholipids, in particular phosphatidylethanolamine (PE) and phosphatidylglycerol (PG), as well as a small fraction of cardiolipin (CL) (45). The outer leaflet faces towards the exterior and is mainly composed of the glycolipid lipopolysaccharide (LPS) (46). LPS consists of three separate regions: a hydrophobic lipid A tail that anchors the LPS molecule to the OM, a phosphorylated core oligosaccharide (OS) and a repeating OS unit which is known as the O-antigen polysaccharide (Figure 6). Figure 6. Structural composition of LPS from . The LPS molecule can be structurally divided into three parts: a lipid A part which embeds the LPS molecule into the OM, a

core OS which consists of a conserved inner core and a less conserved outer core, and an O-antigen which is the outermost part of the LPS molecule. LPS is a heterogeneous molecule and can have various modifications in each part (47). glycero-heptose; Glu, glucose; Gal, galactose; P, phosphate; PEtN, phosphoethanolamine; Kdo, 3-deoxy-manno-oct-2-ulosonic acid; GlcN, glucosamine. Adapted from (48). Reprinted with permission. The lipid A part (also known as endotoxin) is highly hydrophobic and activates the adaptive immune response (49). It consists of a phosphorylated diglucosamine backbone with covalently linked acyl chains that anchor the LPS molecule into the OM. The structure of lipid A varies among bacterial species and modification of the lipid A component can affect the endotoxic activity (47). For example, lipid A modifying enzymes affect the number, position and length of acyl chains as well as the phosphorylation pattern or the sugar composition of the disaccharide backbone (50). The core OS is attached to the lipid A and divided into two parts: the inner core is largely conserved in Gram-negative bacteria and composed of the sugars 3-deoxy--oct-2-ulosonic acid (Kdo) and The outer core is less conserved and composed of different hexose sugars. Both the inner and outer core can be modified by the addition of phosphates, pyrophosphates, various phospholipids and amino acids (47). The outermost part of the LPS molecule is the O-antige

n which is linked to the core OS. The O-antigen is the major antigen targeted by the host immune response and consists of repeating units of oligosaccharides with high variability (51). Each of these repeating units contains one to eight glycosyl residues and varies between sequence, chemical linkage and sugar composition. The OM is also spanned by outer membrane proteins (OMPs). OMPs have a -barrel structure and mediate transport, signalling and other vital functions (52). Some of these OMPs form water-filled channels which are also known as porins. Porins regulate the uptake of nutrients, ions and other molecules such as antibiotics (53, 54). The OM also contains lipoproteins, which are anchored through an N-terminal lipid moiety to the inner leaflet of the OM. In play important roles in the assembly of the OM (55). The most abundant lipoprotein in is Lpp with more than 500,000 copies per cell (43). Lpp provides structural stability to the cell envelope by covalently linking the OM with the underlying peptidoglycan layer (Figure 5) (56, 57). 28 1.5.2 The periplasmic space The periplasmic space is an aqueous, protein-filled environment that lies between the IM and OM (Figure 5). This compartment contains various proteins and polysaccharides that regulate cellular processes such as nutrient uptake, detoxification of harmful compounds or protein transport and quality control (58). It also includes a thin layer of peptidoglycan that stabiliz

es the cell membranes against internal osmotic pressures (43). The peptidoglycan layer is composed of linear glycan strands that consist of alternating units of -acetylglucosamine and -acetylmuramic acid, which are connected by (1,4)-glycosidic bonds (59). The glycan strands are crosslinked by short peptide linkages which create a net-like polymeric structure. The IM separates the periplasm from the cytosol and is composed of two leaflets that form a symmetric phospholipid bilayer. The phospholipid composition of the IM is comparable to the inner leaflet of the OM. In E. , the IM consists of 70 … 80 % PE, 15 … 20 % PG and a small fraction of CL (60, 61). The IM is composed of various proteins that mediate essential cellular functions such as selective transport, energy metabolism, cell division, motility, and signalling (62). IM proteins are generally divided into two classes: integral membrane proteins contain one or more transmembrane segments that span the phospholipid bilayer. Peripherally attached membrane proteins do not span the IM but instead adhere to one of the leaflets via electrostatic, hydrophobic or non-covalent interactions. Another group of proteins which interacts with the IM are lipoproteins. Lipoproteins are anchored to the outer leaflet of the IM through their N-terminal lipid moieties and involved in various processes of the bacterial cell envelope (63). 1.6 Antibiotic transport across the Gram-negative All antibiotics

need to cross one or both membranes of the cell envelope in order to reach their intracellular target. In addition to the physical barriers provided by the bacterial membranes, multiple efflux pumps reduce the intracellular antibiotic concentration. The first and most relevant barrier that antibiotics encounter is provided by the OM (53, 64). Antibiotics have two choices to cross this OM: they can either diffuse through water-filled porin channels or permeate through the LPS-containing leaflet of the OM. In this section, I will address how the various barriers of the cell envelope limit the efficient uptake of certain antibiotics in Gram-negative bacteria. 1.6.1 Porin-mediated antibiotic uptake The uptake of hydrophilic antibiotics (herein termed as small-scaffold antibiotics) is largely regulated by water-filled channels known as porins (Figure 7) (53, 54, 65). Porins can be classified into different groups according to their structure, regulation, expression and activity (66). General diffusion porins form trimers of 16-stranded of hydrophilic molecules with limited substrate selectivity (54, 64). In contrast, substrate-specific porins form trimers of 18-stranded -barrels and are selective for specific substrate classes such as higher oligosaccharides (53). Most porins which are involved in antibiotic transport belong either to the OmpC or OmpF subfamilies (54). These porins represent the main entry pathway for small, hydrophilic antibiot

ics such as -lactams or fluoroquinolones (54, 67). has three major diffusion porins, namely OmpC, OmpF and PhoE. OmpC and OmpF porins have a slight preference 30 for cations whereas PhoE prefers inorganic phosphate and anions (53, 65). Other factors such as molecular shape, number of rotatable bonds and the presence of an ionisable nitrogen group also affect diffusion across porin Bacterial cells have different strategies to acquire resistance to antibiotics which diffuse through porin channels. This can be caused by exchanges in the porin type, changes in the expression profile of porins or by mutations or modifications which affect the channel properties of porins (54). For example, a study showed that clinical isolates of K. pneumonia had a reduced OM permeability after antibiotic treatment due to a change in the expression of porins from large channel size OmpF to the smaller channel size OmpC (69). In addition, some Gram-negative bacteria such as possess a high intrinsic resistance towards many antibiotics (e.g. -lactams) which is partly caused by the low number of general diffusion porins (53). Antibiotics larger than 600 Daltons (herein called large-scaffold antibiotics) do not efficiently cross porin channels (Figure 7). They can permeate through the LPS-containing OM but their diffusion is often limited. It has been shown that lipophilic molecules permeate much less efficiently across the OM bilayer than through a standard phospho

lipid bilayer (70, 71). In addition, the LPS-containing bilayer largely restricts the diffusion of hydrophilic molecules. Hence, the OM is an effective barrier against both hydrophobic as well as hydrophilic molecules (53, 72). The impermeability of the OM can be mainly attributed to the LPS layer. This layer is stabilized by divalent cations, such as Mg cross-bridge negatively charged LPS molecules and thus stabilize the lateral interaction between neighbouring LPS molecules (53). Displacement of the divalent cations by the chelator EDTA destabilizes the LPS layer (73). Subsequently, LPS is released into the medium and it is assumed that phospholipids from the inner leaflet compensate for the loss. As a consequence, Gram-negative bacteria become sensitized towards hydrophobic antibiotics including erythromycin, rifampicin and novobiocin (74). The core region of LPS plays an important role and contributes to the barrier function (64). Strains expressing full-length LPS (termed smooth LPS) are intrinsically resistant to hydrophobic and large-scaffold antibiotics. However, strains expressing truncated LPS (termed rough or deep rough LPS) are more susceptible (64, 75). Deep rough mutants which have the most truncated core are also highly susceptible to lipophilic agents such as detergents, bile salts and some antibiotics (64). 1.6.3 Diffusion across the periplasmic space and IM Antibiotics that have crossed the OM arrive in the periplasmic spa

ce. This compartment is the site of action for antibiotics such as the -lactams or the glycopeptide antibiotic vancomycin. Little is known about the diffusion of antibiotics through this compartment but it is believed that the periplasm represents no major barrier for antibiotics (76). Those antibiotics destined for the cytosol need to penetrate the IM. The phospholipid bilayer of the IM is largely permeable to lipophilic antibiotics (67, 77). Thus, relatively hydrophobic antibiotics such as the macrolides, lincosamides, oxazolidinones, pleuromutilins streptogramins A and B, the elfamycins, and rifamycins appear to cross the IM by simple diffusion (78). In contrast, the IM is largely impermeable to large, uncharged polar molecules and all 32 charged molecules including ions (79). Those antibiotics which belong to these categories need to cross the IM by specific uptake systems. For example, D-cycloserine is transported across the IM via the D-alanine transport system and coupled to the proton motive force (80). Another example is fosfomycin which is transported across the IM by using the glycerol-3-phosphate or the hexose phosphate transporters(81). Transport of aminoglycosides requires both the electrochemical gradient across the IM and the electron flow through the respiratory chain (82). Despite the characterization of a few transport systems, antibiotic transport across the IM remains poorly characterized. Notably, it has also been sugg

ested that all hydrophilic antibiotics traverse the IM by solute transport systems (83). Figure 7. Antibiotic uptake in Gram-negative bacteria. Small-scaffold antibiotics (600 Da) generally permeate the OM by passively diffusing through non-specific OMPs. In contrast, large-scaffold &#x 000;antibiotics ( 600 Da) diffuse through the LPS-containing OM to gain access to the cell interior. However, their diffusion is limited and thus their uptake inefficient. Both large and small scaffold antibiotics can either diffuse across the IM or they can be inadvertently taken up by membrane-embedded transporters. 1.6.4 Efflux pumps Efflux pumps are conserved in nearly all bacterial species and the genes encoding for efflux pumps are either located on the bacterial chromosome or on transmissible elements such as plasmids (84). Bacterial efflux pumps can be classified into several families based on the number of components, the number of transmembrane-spanning regions and the energy source they use to transport their substrate (85). Efflux pumps can be specific for a substrate or they can expel a broad range of structurally diverse compounds (86). There are five families of efflux pumps which are associated with multidrug resistance (MDR): these include the ATP-binding cassette (ABC) superfamiliy, the major facilitator superfamily (MFS), the multidrug and toxic compound extrusion (MATE) family, the small multidrug resistance (SMR) family and the resist

ance-nodulation-division (RND) family (85). The ABC family transporters use the energy of ATP-hydrolysis for drug export whereas the others are antiporters and dependent on the Hproton gradient (87). The major relevant efflux pumps in Gram-negative bacteria belong to the RND-type family which is often associated with multidrug resistance in clinical isolates (67). Efflux pumps of the RND-type family are organised as tripartite systems which are composed of an IM transporter, a periplasmic adapter protein and an OM protein. RND-type efflux systems are polyspecific and expel a broad range of substrates (88). Two of the best studied RND-type efflux pumps in Gram-negative bacteria are the AcrAB-TolC system from E. coli and the MexAB-OprM system from . The AcrAB-TolC efflux pump from is constitutively expressed and has a broad substrate profile which includes antimicrobials such as fluoroquinolones, lipophilic -lactams, chloramphenicol, rifampicin, novobiocin, tetracycline and fusidic acid (85, 89, 90). Several Acr efflux pumps have been chraracterized in E. coli (91…93). However, AcrAB-TolC has been found to be the most relevant one in clinical isolates (94…96). In , several RND-type efflux pumps 34 have been characterized among which MexAB-OprM is the major multidrug efflux system (97). MexAB-OprM is consitutively expressed in wild type and involved in the transport of various antimicrobials -lactams, macrolides, tetracyclines, trimethoprim,

sulfamides and chloramphenicol (98). MexAB-OprM contributes significantly to the high intrinsic resistance and overexpression of it has been associated with MDR in clinical isolates of P. aeruginosa (99). As previously stated, the development of novel antibiotics is not sufficient to keep up with the increasing number of antibiotic resistant bacteria. Thus, there is a pressing need to explore alternative strategies to tackle the issue. One alternative strategy which has received increased attention over the past few years is the development of adjuvants to potentiate the activity of existing antibiotics (100…102). Antibiotic adjuvants are compounds which normally have little or no intrinsic antimicrobial activity. However, when used in combination with an antibiotic, adjuvants enhance the activity of the antibiotic. Antibiotic adjuvants can have different cellular targets but they act by either reversing acquired resistance or sensitizing strains that are intrinsically resistant (100). Most adjuvants that have been identified potentiate antibiotics by either inhibiting antibiotic resistance elements (e.g. -lactamases), blocking bacterial efflux pumps, increasing the permeability of the bacterial cell envelope, or by interfering with signal systems that are involved in antibiotic resistance (102). In this section, I would like to focus on a few antibiotic adjuvants which belong to one of these categories and have been shown to potentiate an

tibiotics in Gram- -lactamase inhibitors are clinically used to improve the activity of -lactam antibiotics. One of the most successful drugs is Augmentin which is a combination of clavulanic acid (a -lactamase inhibitor) and amoxicillin (a -lactam antibiotic). Augmentin is used to restore or extend the antimicrobial activity of amoxicillin in -lactamase-producing strains such and K. pneumoniae (103). Despite the clinical use of Augmentinfor more than 30 years, the emergence of resistance in clinical isolates has been very low (104). Potentiation of antibiotics by inactivation of currently in development (105). Efflux Pumps are another target to potentiate the activity of antibiotics. Since most antibiotics are susceptible to active efflux, the use of efflux pump inhibitors (EPIs) could make antibiotics more effective by increasing their intracellular concentration. A number of EPIs have been identified in Gram-negative bacteria (106, 107). One prominent example is the peptidomimetic compound phenylalanine-arginine -naphthylamide (PAN), which inhibits the Mex efflux pumps in and the homolog AcrAB-TolC in (108). PAN potentiated the activity of the fluoroquinolone antibiotic (108). In particular, PAN reduced the MIC (Minimal Inhibitory Concentration) of levofloxacin about 8-fold in WT strains and up to 64-fold in strains overexpressing efflux pumps. PAN also decreased the MICs of various antibiotics including chloramphenicol, nalidixic ac

id, ofloxaxin, cloxaxillin and erythromycin in K. pneumoniastrains that are either naturally sensitive or inherently resistant against these activity against clinically relevant multidrug efflux pumps, PAN and its derivatives did not enter into clinics due to acute toxicity in pre-clinical trials (110). example, the OM permeabilizer polymyxin B nonapeptide (PMBN) has been shown to potentiate various antibiotics in Gram-negative bacteria (74). 36 PMBN improved the activity of erythromycin and novobiocin in mice infected with either or (111). A combination of PMBN with either avibactam, cefazidime or cefazidime-avibactam increased the antimicrobial activity against clinical isolates of (112). However, despite good in vitro activity against Gram-negative bacteria, PMBN has been shown to cause a similar toxicity profile as polymyxin B (113). Of recent interest are the findings of Wright and colleagues (114): they screened a collection of previously approved drugs for potentiators of the tetracycline antibiotic minocycline against P. aeruginosa and . The authors identified several non-antibiotic compounds which had antibiotic adjuvant properties. For example, the anti-diarrheal medication loperamide improved the activity of minocycline and other tetracycline . Loperamide dissipates the electrical component of the proton motive force across the bacterial membrane of Gram-negative bacteria. As a consequence, loperamide increases antibiotic i

nflux of tetracycline antibiotics. The combination of loperamide and minocycline was shown to be highly synergistic in a model. The same group performed an additional screen to identify antibiotic adjuvants of the aminocoumarin antibiotic novobiocin, which is normally ineffective against Gram-negative bacteria (115). They identified four compounds that were shown to be synergistic with novobiocin in either P. aeruginosa. Two of these compounds, namely A22 and pivmecillinam alter bacterial cell shape by blocking the cytoskeleton protein MreB and inhibiting peptidoglycan synthesis, respectively. The number of antibiotic adjuvants is most likely not sufficient to solve the antibiotic crisis. However, recent genetic studies have provided novel targets for antibiotic adjuvants by identifying a large number of genes that contribute to the intrinsic resistance of Gram-negative bacteria. For example, a transposon mutant library has been used to identify genes that are (116). Other studies have used transposon mutant libraries of to screen for enhanced sensitivity against tobramycin or ciprofloxacin (117, 118). In addition, Fajardo screened two different transposon-tagged insertion libraries of for increased susceptibility to six antimicrobials belonging to different structural families (119). Similar studies have also been done in . For example, Tamae . and Liu screened an knockout collection to look for mutants which are more susceptible t

o various antibiotics (75, 120). Taken together, these genetic studies have identified a large number of additional targets that could be inhibited by adjuvants to improve the activity of various antibiotics. . who screened an knockout collection of close to 4,000 strains, each lacking a different non-essential gene (75). Their goal was to identify strains that were more susceptible to a panel of 22 antimicrobials. In doing so, the authors determined an antibiotic susceptibility profile for each mutant. Out of these 4,000 strains, they identified 283 that showed enhanced sensitivity to at least one or more of the 22 tested antibiotics. These strains could be classified into 7 different categories based on the cellular function of the deleted gene 38 Figure 8. Functional categorisation of the K12 deletion mutants that are hypersensitive to antibiotics. 283 E. coli strains showed increased sensitivity to at least one or more of the 22 antibiotics. The strains were grouped into 7 different categories based on the cellular function of the deleted gene. These include category 1: DNA replication, recombination and repair, category 2: transport, efflux, cell wall and cell membrane synthesis, category 3: protein synthesis, category 4: central metabolic reactions, category 5: regulation, category 6: prophage-carried genes and cell adhesion, and category 7: unassigned gene products. Data taken from (75). Reprinted with Interestingly, 96 of the 283

strains (34 %) could be classified into the category which is involved in transport, efflux, cell wall and cell membrane synthesis (Figure 8), indicating that components of the cell envelope restrict the action of many antibiotics. Table 1 shows the sensitivity profiles of each of the 96 strains that belong to this category. Most of these mutants showed enhanced sensitivity to several antibiotics. Strikingly, some of these mutants or were more susceptible to nearly all of the tested antibiotics (Table 1). The products of the two genes are components of the multidrug efflux pump AcrAB-TolC. Namely, AcrA is the periplasmic adapter protein of AcrAB-TolC whereas TolC represents the 40 Table 1. Antibiotic sensitivity profiles of the 96 E. coli K12 knockout strains that are involved in transport, efflux, cell wall and cell membrane synthesis. Sensitivity profiles were categorized into three different groups with strong susceptibilities in dark shades, medium susceptibilities in lighter shades and weak susceptibilities in lightest shade. CIP, Ciprofloxacin; ENX, Enoxacin; NIT, Nitrofurantoin; MTR, Metronidazole; SFX, Sulfamethoxazole; RIF, Rifampicin; GEN, Gentamicin; TOB, Tobramycin; NEO, Neomycin; STR, Streptomycin; SPT, Spectinomycin; TET, Tetracycline; VAN, Vancomycin; AMP, Ampicillin; RAD, Cephradine; FOX, Cefoxitin; ATM, Azetreonam, CST; Colistin; CHL, Chloramphenicol; ERY, Erythromycin; FUS, Fusidic acid; TRI, Triclosan. Large-scaff

old antibiotics rifampicin, vancomycin and erythromycin are boxed in red. Data taken from Liu et al (75). Reprinted with permission. Since one of the goals of my doctoral thesis was to improve the activity of those antibiotics which do not efficiently cross the OM, I was particularly interested in the mutants that showed increased susceptibility to the large-scaffold antibiotics rifampicin, vancomycin and erythromycin (Table 1). Initially I focused on the glycopeptide antibiotic vancomycin. Vancomycin is largely ineffective against Gram-negative bacteria because it does not efficiently penetrate the OM. However, as shown in Table 2, there are 60 E. coli whose inactivation improves the activity of vancomycin. Notably, some of these mutants have previously been shown to have increased sensitivity towards vancomycin. For example, the hypersensitive mutant was 125-fold more sensitive to vancomycin than the corresponding WT strain (Table 2). In addition, the mutant became also sensitized to other large-scaffold antibiotics including rifampicin and erythromycin (Table 1). SurA is a periplasmic chaperone that is involved in the folding and transport of periplasmic proteins. Strains lacking SurA are defective in the assembly of the OM, which results in an increased sensitivity towards vancomycin and other antimicrobials (121, 122). Other mutants which have been previously described to be more sensitive towards vancomycin include and (Table

2). They encode for the lipoproteins BamB and BamE which are part of the -barrel assembly machinery) complex. This complex catalyses the folding and insertion of OMPs and is essential for the integrity of the OM (123). Mutants lacking either BamB or BamE have a severe defect in the OM which results in an increased sensitivity to various antibiotics including vancomycin (124, 125). 42 Gene MIC g/ml) Protein name BW25113 wt 500 4 2 Peptidyl-prolyl cis-trans isomerase SurA smpA 70 2 OM protein assembly factor BamE bamB 100 2 OM protein assembly factor BamB envC 100 2 Murein hydrolase activator EnvC lpxL 100 2 Lipid A biosynthesis lauroyl acetyltransferase tatC 100 2 Twin arginine protein translocation system tolR 100 2 Colicin transport; Tol-Pal system component ydcS 100 2 Putative ABC transporter periplasmic-binding protein YdcS yciM Lipopolysaccharide assembly protein B recA DNA strand exchange and recombination protein envZ 150 2 Osmolarity sensor protein EnvZ lpxM 150 2 Lipid A biosynthesis myristoyltransferase nlpC 150 2 Probable endopeptidase NlpC pal 150 2 Peptidoglycan-associated lipoprotein Pal proW 150 2 Glycine betaine/L-proline transport system permease protein ProW rfaC 150 2 Lipopolysaccharide heptosyltransferase I ybgF 150 2 Periplasmic TolA-binding protein yhdP 150 2 Uncharacterized protein YhdP dnaK 150 2 Chaperone protein DnaK hlpA 150 2 Periplasmic chaperone Skp hscA 150 2 Iron-su

lfur cluster biosynthesis chaperone HscA hscB 150 2 Co-chaperone for iron-sulfur cluster biosynthesis rimK Ribosomal protein S6 modification protein rlpA Endolytic peptidoglycan transglycosylase RlpA tufA Translation elongation factor Tu 1 yfgC -barrel assembly-enhancing protease aceE Pyruvate dehydrogenase E1 component pgaC Poly--1,6-N-acetyl--glucosamine synthase ycjU -phosphoglucomutase ygcO Ferredoxin-like protein YgcO dksA RNA polymerase-binding transcription factor DksA fur Ferric uptake regulation protein rsmF Ribosomal RNA small subunit methyltransferase F xapR HTH-type transcriptional regulator XapR yciT Uncharacterized HTH-type transcriptional regulator YciT ylcG Uncharacterized protein YlcG ydhT Uncharacterized protein YdhT recB � 150 1 RecBCD enzyme subunit RecB ftsP � 150 1 Cell division protein FtsP fepC � 150 2 Ferric enterobactin transport ATP-binding protein FepC qmcA � 150 2 Protein QmcA tatB � 150 2 Sec-independent protein translocase protein TatB tonB � 150 2 Ton complex subunit TonB yheL � 150 2 Sulfurtransferase complex subunit TusB elaD � 150 3 Protease ElaD rpmJ � 150 3 50S ribosomal protein L36 rpsO � 150 3 30S ribosomal protein S15 � 150 3 Ribosomal RNA large subunit methyltransferase E Table 2. List of K12 deletion strains which showed increased sensitivity to vancomyci

n. Category 1: DNA replication, recombination and repair, category 2: transport, efflux, cell wall and cell membrane synthesis, category 3: protein synthesis, category 4: central metabolic reactions, category 5: regulation, category 6: prophage-carried genes and cell adhesion, and category 7: unassigned gene products. Data taken from (75). Reprinted with permission. Interestingly, most of the remaining mutants in Table 2 had not been previously described to show increased sensitivity towards vancomycin. Notably, some of these mutants such as lpxM or are involved in LPS biosynthesis. Both LpxL and LpxM are acyltransferases which transfer either a laurate or myristate chain onto the Kdo-lipid A precursor (126). RfaC (also known as WaaC) is a heptosyltransferase which transfers the first heptose sugar onto the Kdo moiety of the LPS inner core (127). YciM has been recently characterized as a modulator of LPS levels by negatively regulating the biosynthesis of lipid A (128). Based on the findings of this study, we reasoned that a small molecule which inhibits any of the targets listed in Table 2 could be used as an antibiotic adjuvant to enhance the activity of vancomycin. Since many vancomycin-sensitive strains are also more susceptible to other antibiotics (Table 1), we believe that such a small molecule inhibitor could potentially be used to enhance the activity of several antibiotics. Hence, we carried out a high-throughput screen in p

aper III to identify such small molecules. yheM � 150 3 Protein TusC yheN � 150 3 Sulfurtransferase TusD rppH � 150 3 RNA pyrophosphohydrolase cls � 150 4 Cardiolipin synthase A cydB � 150 4 Cytochrome bd-I ubiquinol oxidase subunit 2 fdx � 150 4 2Fe-2S ferredoxin ytjC � 150 4 Probable phosphoglycerate mutase GpmB gpmM � 150 4 2,3-bisphosphoglycerate-independent phosphoglycerate mutase iscS � 150 4 Cysteine desulfurase IscS hns � 150 5 DNA-binding protein H-NS ybgT � 150 7 Cytochrome bd-I ubiquinol oxidase subunit X yjjY � 150 7 Uncharacterized protein YjjY 44 The overarching goal of my thesis was to understand how the OM barrier of could be weakened. One approach that I took was to investigate the feasibility of using small molecules to increase the permeability of the OM. We reasoned that small molecules which permeabilize the OM could be used as adjuvants or lead molecules for adjuvants to improve the activity of large-As a prelude to this, I initially investigated the periplasmic chaperone network which is required for the biogenesis of OMPs. (Figure 9, left panel). OMPs are synthesized on cytosolic ribosomes as precursors with an N-terminal signal sequence and then transported across the IM by the SecYEG translocase (129). After translocation and cleavage of the signal sequence, newly exported OMPs are transported from the IM

to the OM by the periplasmic chaperone network SurA, Skp and DegP (130). The periplasmic chaperone network, in particular SurA and Skp, is essential for the integrity of the OM and strains lacking either of them are more sensitive to various large-scaffold antibiotics including vancomycin (121, 122, 131). Figure 9. Three major pathways required for the biogenesis of the OM in Gram-negative bacteria. Left panel: OMPs are synthesized on cytosolic ribosomes and subsequently translocated across the IM by the SecYEG translocase. Unfolded OMPs are trafficked across the periplasmic space by the periplasmic chaperone network SurA, Skp and DegP, and delivered to the -barrel assembly machinery (BAM) complex, which folds and integrates OMPs into the OM. Middle panel: Lipoproteins are trafficked across the IM via the SecYEG translocase and subsequently processed at the outer leaflet of the IM. Processed lipoproteins destined to the OM are extracted from the IM by the ABC transporter LolCDE, and then transferred to the periplasmic chaperone LolA. LolA traffics lipoproteins across the periplasmic space to the OM lipoprotein LolB which incorporates lipoproteins into the OM. Right panel: Lipopolysaccharide synthesis occurs at the inner leaflet of the IM. Nascent LPS molecules are translocated across the IM by the ABC transporter MsbA. Subsequently, LPS molecules are extracted from the IM by the LptBCFG complex and then translocated across the periplasmic s

pace via a transenvelope bridge. Once arrived at the OM, LPS molecules are integrated into the OM by the LptD-LptE complex. Adapted from (132). Reprinted with permission. 46 Before I joined the Daley lab, my former coworker Jörg Götzke identified a protein with no annotated function, namely YfgM, which we thought might be a novel member of this periplasmic chaperone network. He was able to show that YfgM forms a complex with the periplasmic chaperone PpiD at the SecYEG translocon (Figure 9, left panel). At this point, the function of YfgM remained unknown. PpiD has been characterized as a member of the periplasmic chaperone network, which assists in the release of newly translocated OMPs proteins by preventing their premature aggregation (133, 134). Since YfgM interacts with PpiD at the SecYEG translocon, we reasoned that YfgM might also be a part of this chaperone network. , we had a closer look at the bacterial chromosome and found that is located upstream of . As previously mentioned, BamB is a subunit of the BAM complex which is required for the integrity of the OM (Figure 9, left panel). Strains lacking BamB have a defect in their OM and thus are susceptible towards vancomycin. Hence, I was interested to investigate if strains lacking YfgM also have a defect in their OM. To address this question, we engineered strains that were lacking YfgM, PpiD or other periplasmic chaperones such as SurA, Skp or DegP. We then analyzed the integrity

of the OM in these strains by performing an antibiotic disc diffusion assay. In this assay, a filter disk containing a sub-inhibitory concentration of vancomycin was spotted onto a lawn of bacteria (Figure 10A). Strains which have an intact OM, such as WT resistant towards vancomycin. Hence, no growth inhibition zone is observed around the filter disk (Figure 10A, top panel). In contrast, strains with a , have a compromised OM. In these strains, vancomycin can access its target in the periplasmic space, which results in growth inhibition around the filter disk (Figure 10A, lower Figure 10. Disc diffusion assay to investigate the integrity of the OM. (A) WT E. colistrains are intrinsically resistant towards vancomycin. Hence, no zone of growth inhibition is observed around the filter disk (top panel). Strains with a defect in their OM, such as the strain, became sensitized towards vancomycin which results in a growth inhibition zone around the filter disk (bottom panel). (B) Disc diffusion assays were performed in WT and various deletion strains to investigate the integrity of the OM. Figure 10B taken from (135). Reprinted with permission. Our data showed that deletion of the gene encoding for YfgM did not cause any obvious defects in the OM (Figure 10B). In addition, we observed the same phenotype for strains that were either lacking PpiD or DegP (Figure 10B). However, as previously reported, deletion of either SurA or Skp destabilized

the OM, which increased the sensitivity towards vancomycin Since periplasmic chaperones often have overlapping functions, their role in the periplasmic chaperone network can only be elucidated by deleting them in combination. For example, genetic deletion of ppiD and results in a temperature-sensitive phenotype (134). Thus, we investigated if YfgM had overlapping functions with other periplasmic chaperones. To address this question, we engineered double knockout strains that were lacking YfgM in 48 combination with either PpiD, DegP, Skp or SurA. By using the vancomycin susceptibility assay, we investigated the integrity of the OM in these strains and found that deletion of YfgM in strains with either a background did not cause any obvious defects in the OM (Figure 10B). However, deletion of YfgM in strains with either a background further compromised the integrity of the OM, as evidenced by an increased growth inhibition zone around the filter disk containing vancomycin (Figure 10B). Hence, these experiments suggest that YfgM is a novel periplasmic chaperone which operates in the same network as Skp and In a follow-up study function of YfgM by identifying its putative substrates. We hypothesized that substrates of YfgM would be incorrectly folded or trafficked when YfgM was absent from the cell, and therefore more prone to proteolytic degradation. In this study, we used a comparative proteomic approach to quantify the steady-state levels

of proteins in strains lacking . Since our previous study indicated that YfgM has overlapping functions with SurA and Skp, we also included strains into the proteomic analysis that were lacking together with either or , to exclude compensatory effects caused by the major periplasmic chaperones. Surprisingly, the proteomic analysis did not identify any changes in the levels of OMPs or lipoproteins. However, we identified 9 IM proteins and 7 periplasmic proteins whose abundance was significantly changed. One key finding was that a few proteins involved in the adaptation to gastrointestinal stress (e.g. acid-resistance) were lower in abundance in strains lacking YfgM. For example, the periplasmic chaperone HdeB which is involved in the acid-stress response was lower in abundance in all three strains lacking YfgM. To investigate if HdeB was misfolded / mistargeted in strains lacking and thus turned over faster, we performed pulse-chase experiments. In these pulse-chase experiments, we could show that HdeB was turned-over yfgM. The identification of HdeB and other cell envelope proteins as potential substrates of YfgM will be a valuable resource for follow-up experiments that aim to decipher the function of YfgM. Paper I + II provided novel insights into the periplasmic chaperone network and how it contributes to the permeability of the OM. However, the findings of these two studies do not provide any direct application that could be used to

improve the activity of antibiotics. In our goal was to identify a small molecule which causes the OM to be more permeable and thus might be used as an antibiotic adjuvant. Initially, we wanted to identify an inhibitor against the periplasmic chaperone SurA. Hence, we performed a high-throughput screen in which we monitored growth of WT in the presence of a sub-inhibitory concentration of vancomycin (150 µg ml28,000 small molecules. We reasoned that a small molecule that inhibits SurA would increase the activity of vancomycin, which would result in a growth arrest. I established the assay conditions whereas the high-throughput screen was performed by our collaboration partner from the Department of Chemistry at Umeå University. The high-throughput screen identified one promising molecule, namely MAC-13243. This small molecule has been previously identified as an inhibitor of the periplasmic chaperone LolA (136). LolA is an essential protein that functions as a periplasmic shuttle that transports lipoproteins from the IM to the OM (Figure 9, middle panel). All lipoproteins destined for the OM are synthesized in the cytoplasm as precursors with an N-terminal sequence. After translocation through the SecYEG translocon, lipoproteins are processed at the periplasmic side of the IM, which involves sequential modification of a cysteine residue and cleavage of the signal peptide by the lipoprotein-specific signal peptidase LspA (63). Processed l

ipoproteins remain either attached to the IM or they are extracted from the IM by the ABC transporter LolCDE (Figure 9, middle 50 panel). Subsequently, the periplasmic chaperone LolA captures lipoproteins from the LolCDE complex and shuttles them across the periplasmic space to the OM. At the OM, lipoproteins are transferred from LolA to LolB, which localizes them to the OM. Since inhibition of LolA by MAC-13243 leads to decreased levels of OM lipoproteins (136), we hypothesized that this might affect the integrity of the OM. To explore this possibility, I established an NPN (N-phenyl-1-naphthylamine) uptake assay to investigate if the small molecule MAC-13243 could be used to increase the permeability of the OM. The NPN uptake assay is a commonly used method to analyze the integrity of the OM (137). For example, WT strains have an intact OM and thus the hydrophobic dye NPN cannot efficiently cross the OM. However, when the OM is damaged, NPN can access phospholipids in the IM and the inner leaflet of the OM, which results in prominent fluorescence. Using this assay, I demonstrated that MAC-13243, when used at sub-inhibitory concentrations, permeabilizes the OM of WT in a concentration-dependent manner (Figure 11). Figure 11. NPN uptake in WT cells. Left panel: WT E. coli cells were exposed to different concentrations of MAC-13243 and NPN uptake was monitored. Right panel: The increase in fluorescence was considered to be due to increa

sed permeability of the OM since the amount of MAC-13243 did not reduce cell viability. Abbreviations: MIC = Minimal Inhibitory Concentration. Figure taken from paper III. Since MAC-13243 permeabilized the OM of WT E. coli cells, I was curious to investigate if MAC-13243 could also be used as an adjuvant for antibiotics which do not efficiently cross the OM. To address this question, I performed a series of checkerboard assays. The checkerboard assay is a commonly used method to evaluate interactions between two drugs. In brief, I found that MAC-13243 worked synergistically with either erythromycin or novobiocin, meaning that there was added benefit when those two drugs were used in combination (Figure 12). Figure 12. MAC-13243 works synergistically with various large-scaffold antibiotics.Heat plots illustrate inhibition of growth of WT E. coli in the presence of MAC-13243 and vancomycin, rifampicin, erythromycin or novobiocin. Growth percentage of E. coli is shown with different colours where black represents 100% growth and red 0% growth. Figure taken from paper III. At this point, it remained unclear if the leaky phenotype, which we observed cells with MAC-13243, was caused by inhibition of LolA. Since MAC-13243 is not stable in aqueous solution and a structural analogue of its degradation product S-(4-chlorobenzyl)isothiourea, namely A22, has been reported to inhibit the eukaryotic actin-homolog MreB (Figure 17), it remained unclea

r what caused the cells to become more permeable (138). To address this question, I used the CRISPRi system to reduce the intracellular levels of LolA and showed that partial depletion of 52 LolA was sufficient to induce the permeable phenotype, as evidenced by an increased permeability of the OM. To summarize this study, we showed that inhibition of lipoprotein trafficking could be an attractive target for small molecules. By performing a high-throughput screen, we identified an inhibitor of the periplasmic chaperone LolA, namely MAC-13243, and showed that this small molecule can be used as an adjuvant to improve the activity of certain large-scaffold antibiotics. In a parallel approach, we have collaborated with the Widmalm group from the organic chemistry section at Stockholm University. Our common goal was to break the OM permeability barrier by using small molecules which ). Previous to this study, the Widmalm group identified 3 small molecules in a fragment-based screen for inhibitors against the glycosyltransferase WaaG (Figure 13A) (139). These molecular scaffolds, namely L1-L3, have low affinity for WaaG and compete with its natural substrate UDP-Glc for binding. This was of particular interest for us since WaaG is a key enzyme in the synthesis of LPS. The peripherally attached IM protein WaaG adds the first glucose residue to the outer core of the growing LPS molecule (Figure 6). Figure 13. Small molecular fragments L1-L3 and s

chematic representation of the glucose transfer by WaaG. (A) L1, 4-(2-amino-1,3-thiazol-4-yl)phenol; L2, 4-(1H-pyrrol-1-yl)benzoic acid; L3, 2-(1H-pyrrol-1-ylmethyl)pyridine; (B) The glycosyltransferase WaaG transfers a C-glucose residue from UDP-Glc to the LPS acceptor molecule. Figure 13A taken from (142). Reprinted with permission. WaaG is essential for the stability of the OM and deletion of the gene encoding for WaaG results in an inability to synthesize the outer core and the become more sensitive to several antibiotic classes (Table 1, formerly known ) (75). We hypothesized that a small molecule that inhibits WaaG could be used as an antibiotic adjuvant. Thus, in paper IV I explored if any of these small molecular scaffolds could be used to inhibit WaaG developed an activity assay for WaaG using C-labeled UDP-glucose and LPS that had been purified from a Figure 14. Molecular dynamics simulations of OmpF trimer intercalated in OMs of E. OmpF trimer was intercalated in either E. coli E. coli K12 core R1 core LPS with five repeating units of O6-antigen (right panel). Lipid A is illustrated as pink spheres. Core sugars (gray) and O-antigen polysaccharides (orange) are illustrated as stick models. The inner leaflet consists of PPPE (blue spheres), PVPG (orange spheres), and PVCL2 (magenta spheres). Ca ions are depicted as cyan small spheres, K ions as green small spheres and Cl ions as magenta small spheres. Abbreviations: PPPE, 1-palmit

oyl(16:0)-2-palmitoleoyl(16:1cis-9)-phosphatidylethanolamine; PVPG, 1-palmitoyl(16:0)-2-vacenoyl(18:1 cis-11) phosphatidylglycerol; PVCL2, 1,10-palmitoyl-2,20-vacenoyl cardiolipin. Figure taken from (141). Reprinted with permission. 54 When I started to develop the assay for WaaG, I initially had problems with its low activity. Since WaaG is a peripherally attached IM protein, I explored if its activity could be improved by the addition of various lipids. After several rounds of optimization, I discovered that I could improve the activity of WaaG by including two membrane lipids, namely PG and CL, and the non-ionic detergent CHAPS to the reaction mixture. Using these optimized in vitro conditions, I showed that one of these small molecular scaffolds, namely L1, is a weak inhibitor of WaaG with an IC of ~ 1 mM (Figure 15). Figure 15. L1 inhibits WaaG Left panel: Mixed micelles consisting of 20 mM CHAPS, 10 mM PG and 1 mM CL were added to LPS- and UDP-Glc. The reaction was initiated by adding WaaG and incubated either with 2.5% DMSO or 25 mM of ligand L1, L2 or L3. Samples were collected after various time points and inactivated by using Laemmli buffer. Then, samples were separated by SDS-PAGE and detected by autoradiography. Right panel: Relative activity of WaaG in the presence of either DMSO or 25 mM L1, L2 or L3. Figure taken from (142). Reprinted with permission. The observation that L1 could inhibit WaaG was an interesting finding.

Such small molecular fragments with low affinity for their target s from high M to low mM) often represent a good starting point for the design of a high-affinity inhibitor (143, 144). In this process, also known as fragment-based drug design, small molecular fragments with low affinity for their target are chemically elaborated or linked to produce a high affinity inhibitor. Since L1 could be used to inhibit WaaG chemical space around L1 (and also L2 - L3), and created a fragment-based library that included an additional 17 small molecular fragments (Figure 16, unpublished data). At this point, the aim was to identify additional small molecular fragments that either compete with the natural substrate UDP-Glc for binding or inhibit WaaG in the in vitro activity assay.Figure 16. Small molecular scaffold library L1-20. 56 Hence, I tested all small molecular fragments in the activity assay to evaluate their inhibitory activity. Surprisingly, the results indicated that none of the additional fragments had inhibitory activity (data not shown). These findings were surprising since some of these fragments are structurally closely related to L1 (Figure 16). The Widmalm group is currently investigating by NMR spectroscopy if any of the additional small molecular fragments compete with the natural substrate UDP-Glc for binding. Our preliminary results indicate that one of these fragments, namely L8, also binds to WaaG (data not shown). Despite t

he fact that the expanded library did not contain a more potent inhibitor than L1, we could identify at least one additional small molecular fragment that binds to WaaG. This provides further insight for the design of a potent inhibitor against WaaG. Gram-negative bacteria have developed sophisticated mechanisms to protect themselves against noxious molecules such as antibiotics. Two of the major mechanisms that limit the activity of many antibiotics include active efflux by efflux pumps and reduced uptake across the OM barrier. The presented doctoral thesis had the objective to investigate how the reduced uptake across the OM barrier could be improved by destabilizing the OM. To address this problem, I initially started investigating the periplasmic chaperone network and how it contributes to the permeability of the OM In this paper, we identified a novel component of the SecYEG translocon, namely YfgM, and showed that it operates in the same pathway as the periplasmic chaperone network SurA/Skp. However, YfgM plays only a minor role in this network since strains lacking YfgM did not have any obvious defects in the OM. The molecular function of YfgM remains to be determined but we speculate that it might act as a docking platform for the periplasmic chaperones SurA/Skp. Interestingly, the periplasmic domain of YfgM contains tetratricopeptide repeat (TPR) domains whose function has not been determined yet. These domains are often involved

in protein-protein interactions and thus it might be interesting to further investigate the function of the TPR domains in YfgM (145). There is also evidence that Skp interacts with OMPs during their early translocation through the SecYEG translocon (146). Hence, it is possible that Skp might dock to the SecYEG translocon or a protein in close vicinity to it. To better understand the function of YfgM, we performed a comparative proteomic approach to identify potential substrates of YfgM (). We hypothesized that strains lacking YfgM might have a changed OMP profile 58 since YfgM operates in the same network as SurA/Skp. Although we did not observe any significant changes in the levels of OMPs and lipoproteins in , the proteomic data revealed an unexpected insight into the physiological role of YfgM. We identified a number of proteins that are involved in acid-stress response, which were lower in abundance in strains shown to be induced during acid-stress previously (147…150). In an additional experiment, we could confirm that strains devoid of YfgM had a decreased survival rate at low pH. At this stage, it remains unclear if the decreased survival rate at low pH in strains lacking YfgM is caused by a defect in the cell envelope or by another secondary effect. To address the reduced uptake of antibiotics across the OM, we performed a high-throughput screen to look for small molecules that destabilize the OM ). In this study, we identified an

inhibitor of the periplasmic chaperone LolA, named MAC-13243. This small molecule had been previously identified as a novel antimicrobial lead (136). However, I could show that MAC-13243 can also be repurposed to an antibiotic adjuvant. I observed that MAC-13243 synergized with the large-scaffold antibiotics novobiocin and erythromycin, but not with vancomycin and rifampicin. At this point, it remains unclear to us why MAC-13243 works synergistically with some large-scaffold antibiotics but not with others. It is worth noting that Krishnamoorthy . reported that the activity of these 4 antibiotics is significantly limited by the OM and/or by active efflux (72). Surprisingly, they found that the OM presented no major obstacle for novobiocin whereas active efflux drastically reduced its activity. In the case of rifampicin and vancomycin, they found that the OM barrier significantly limited their activity whereas inactivation of active efflux only minor improved their activity. In the case of erythromycin, a combination of both reduced uptake and active efflux limited its activity. Despite the preliminary findings, I believe that MAC-13243 has certain limitations in its current form that prevent it from being used as an antibiotic adjuvant. First of all, the synergistic interactions between MAC-13243 and either novobiocin or erythromycin were moderate and only observed under specific conditions. One of the major limitations of MAC-13243 migh

t be its limited stability in aqueous solution (151). MAC-13243 degrades in aqueous solution into one molecule of 3,4-dimethoxyphenethylamine, two molecules of formaldehyde and one molecule of S-(4-chlorobenzyl)isothiourea (Figure 17A). Thus, an attempt to improve the chemical stability of MAC-13243 could improve its activity as an antibiotic adjuvant. This could be achieved by stabilizing the central triazine ring of MAC-13243, which is prone to hydrolysis (151). Figure 17. Degradation of MAC-13243 in aqueous solution. (A) MAC-13243 is degraded into one molecule 3,4-dimethoxyphenethylamine, two molecules of formaldehyde and one molecule S(4-chlorobenzyl)isothiourea (151). Both MAC-13243 and the degradation product S(4-chlorobenzyl)isothiourea bind to LolA (151) (B) An analogue of the degradation product, named A22 or S(4-dichlorobenzyl)isothiourea, has been shown to bind to LolA. Interestingly this compound is an inhibitor of the cytoskeletal protein MreB (138). Figure adjusted from (151). Reprinted with permission. 60 Interestingly, one of the degradation products of MAC-13243, namely S-(4-chlorobenzyl)isothiourea, is a structural analogue of compound A22, which has been identified as an inhibitor of the cytoskeletal protein MreB (Figure 17B) (138). Taylor . showed that A22 works synergistically with both novobiocin and rifampicin in (115). Hence, the design of a molecule which inhibits both LolA and MreB could be an interesting concep

t to further enhance the synergistic effects of MAC-13243. I personally think that it would be most interesting to expand on the findings . In this paper, I showed that the small molecular fragment, namely compound L1, can be used to inhibit the glycosyltransferase WaaG . However, due to its weak inhibitory activity, L1 is not a molecule that can be used in its current form as an antibiotic adjuvant. However, we believe that L1 represents a good lead fragment for the design of a more potent inhibitor against WaaG. By expanding chemical space around L1, we could identify an additional molecule, namely L8 (Figure 16), that binds to . In a next optimization round, we plan to systematically evaluate the chemical space around these two molecules to gain further insight into the properties that are required for binding to and inhibiting WaaG. Antibiotika anses vara en av de viktigaste upptäckterna inom humanmedicinen. Sedan de infördes in i klinikerna på 1940-talet har de framgångsrikt använts för att bota bakteriella infektioner. På grund av sin omfattande användning och missbruk under de senaste decennierna har dock många bakteriearter som inte längre är känsliga mot de flesta antibiotika uppstått. Dessa så kallade multiresistenta bakterier har utvecklat olika tillvägagångssätt för att inaktivera antibiotika. Följaktligen har vi ont om behandlingsalternativ och är ofta tvungna att använda äldre antibiotika som är mindre säkra. Resistens mot an

tibiotika är ett allmänt problem. Gramnegativa bakterier är dock arter som är särskilt bekymmersamma. Dessa bakterier dödar tusentals människor årligen och anses därför vara ett stort hot mot folkhälsan. Ett av de största problemen är att många antibiotika är ineffektiva mot gramnegativa bakterier. Detta beror dels på en fysisk barriär, nämligen det yttre membranet, som hämmar effektivt upptag av många antibiotika. I den presenterade doktorsavhandlingen undersökte jag hur dessa fysiska barriärer kunde försvagas. Det ultimata målet med min avhandling var att hitta små molekyler som skulle kunna användas för att bryta denna fysiska barriär. Vi resonerade att dessa små molekyler skulle kunna användas för att förbättra upptagningen av de antibiotika som normalt inte effektivt passerar yttermembranet. 62 I would like to take the opportunity to thank my supervisor for his excellent supervision during the past 5 years. I really enjoyed it to be part of your group. Thanks mate! I also would like to thank our current group members Zoe, and the former group members and It was very pleasant to work with all of you and I think we have a really good ambience in the group. Special thanks to my co-supervisor and his current group members and . It was fun to work together with you guys! Also a special thanks to our former collaborators I am also grateful to my former students Lars Nilsson, Johanna Hultgren, Muna Amanuel, Andreas Dunge and Anja Blümle

r. Thanks to our neighbors from the Jan-Willem de Gier group: ThomasZhe and . Special thanks to my table soccer friend Thomas for never giving up and Alex the Greek for all the insightful philosophical discussions. Thanks to the Rob Daniels group: and the former group members and . It was always nice to see other people in the lab on the weekends. A big thanks to the Gunnar von Heijne and Ingmarie Nilsson group: thanks to my bongo friend Mr. Felix … it was a pleasure to explore Africa with you! Theresa, Genia and Cool group, good people! Thanks to the Elzbieta Glaser group: Pedro and for always being in a good mood and Elzbieta for never-ending but interesting discussions at the lunch table. Many thanks to Brzezinski and Ädelroth group members JacobNathalie and Pia. Great thanks to my climbing and sailing friend Max (aka sailor man) … I knew it was a good We truly had some good laughs and I will never forget the large number of fish we caught - zero! Thanks to the David Drew group members AzizPascal and . Bsunderä Dank äm bärtigä Päsci und sim Schätzi dr Regula - heb sorg zunnärä! Du bisch wahrhaftig en Glückspilz! Sogar dr Andi us Rümlang wär nidisch! Thanks to the Martin Ott group members and for the discussions about the book exam and other Maria SallanderAnn-Britt Rönnell and for running DBB in the background. Last but not least, a very special thanks to my family and friends back home for their great support and visits to Stoc

kholm. Looking forward to see you 64 1. Hall, B. G., Barlow, M. Evolution of the serine -lactamases: past, present and future. 111…123 (2004). 2. Baltz, R. A. Antibiotic discovery from Actinomycetes: Will a renaissance follow the decline and fall? 188…96 (2005). 3. Udekwu, K. I., Parrish, N., Ankomah, P., Baquero, F., Levin, B. R. Functional relationship between bacterial cell density and the efficacy of antibiotics. J. Antimicrob. Chemother. 745…757 (2009). 4. Li, J. Antimicrobial Activity and Resistance: Influencing Factors. Front. Pharmacol.(2017). 5. Scholar, E. M., Pratt, W. B. 2nd edition. Oxford University Press, New York (2000). 6. Lewis, K. Platforms for antibiotic discovery. Nat. Rev. Drug Discov. 371…387 (2013). 7. Kohanski, M. A., Dwyer, D. J., Collins, J. J. How antibiotics kill bacteria: from targets to 423…435 (2010). 8. Kong, K.F., Schneper, L., Mathee, K. -lactam Antibiotics: From Antibiosis to Resistance and APMIS 1…36 (2010). 9. Barlow, M. What antimicrobial resistance has taught us about horizontal gene transfer. Methods 397…411 (2009). 10. Coates, A., Hu, Y., Bax, R., Page, C. The future challenges facing the development of new antimicrobial drugs. Nat. Rev. Drug Discov. 895…910 (2002). 11. Fleming, A. On the Antibacterial Action of Cultures of a , with Special Reference to their Use in the Isolation of 226…236 (1929). 12. Lobanovska, M., Pilla, G. Penicillins Discovery and Antibiotic Resistance: Lessons for the

Future? Yale J. Biol. Med. 135…145 (2017). 13. Coates, A. R., Halls, G., Hu, Y. Novel classes of antibiotics or more of the same? Br. J. 184…194 (2011). 14. Silver, L. L. Challenges of antibacterial discovery. 71…109 (2011). 15. Wenzel, R. P. The Antibiotic Pipeline … Challenges, Costs, and Values. N. Engl. J. Med.351,523…526 (2004). 16. Spellberg, B. The future of antibiotics. 228 (2014). 17. Bakken, J. S. Antibiotic partners promote discovery. Nature 167…167 (2016). 18. Piddock, L. J. V. The crisis of no new antibiotics … What is the way forward? Lancet Infect. Dis. 249…253 (2012). 19. Michael, C. A., Dominey-Howes, D., Labbate, M. The Antimicrobial Resistance Crisis: Causes, Consequences, and Management. Front. Public Health (2014). 20. Payne, D. J., Gwynn, M. N., Holmes, D. J., Pompliano, D. L. Drugs for bad bugs: confronting the challenges of antibacterial discovery. 29…40 (2007). 21. An antibiotic comeback ? Nat. Rev. Drug Discov. 165…165 (2014). 22. Col, N. F., OConnor, R. W. Estimating Worldwide Current Antibiotic Usage: Report of Task Rev. Infect. Dis. S232…S243 (1987). 23. Barbosa, T. M., Levy, S. B. The impact of antibiotic use on resistance development and Drug Resist. Updat. 303…311 (2000). 24. Goossens, H., Ferech, M., Vander Stichele, R., Elseviers, M., ESAC Project Group. Outpatient antibiotic use in Europe and association with resistance: a cross-national database study. Lancet 579…587 (2005). 25. Riedel, S. Antimi

crobial use in Europe and antimicrobial resistance in pneumoniaeEur. J. Clin. Microbiol. 485…490 (2007). 26. Centers for Disease Control and Prevention. Antibiotic Resistance Threats in the United States, 2013. Available at: http://www.cdc.gov/drugresistance/threat-report-2013/index.html, Accessed 20.09.2017 27. Ayukekbong, J. A., Ntemgwa, M., Atabe, A. N. The threat of antimicrobial resistance in developing countries: causes and control strategies. Antimicrob. Resist. Infect. Control 47 (2017). 28. Witte, W. Selective pressure by antibiotic use in livestock. 16, 19…24 (2000). 29. Bengtsson, B., Wierup, M. Antimicrobial Resistance in Scandinavia after a Ban of Antimicrobial Growth Promoters. 147…156 (2006). 30. Butaye, P., Devriese, L. A., Haesebrouck, F. Antimicrobial Growth Promoters Used in Animal Feed: Effects of Less Well Known Antibiotics on Gram-Positive Bacteria. Clin. Microbiol. Rev. 175…188 (2003). 31. Larsson, D. G. J. Pollution from drug manufacturing: review and perspectives. Phil. Trans. R. (2014). 32. Chemaly, R. F. The role of the healthcare environment in the spread of multidrug-resistant organisms: update on current best practices for containment. 79…90 (2014). 33. Sydnor, E. R. M., Perl, T. M. Hospital Epidemiology and Infection Control in Acute-Care 141…173 (2011). 34. The evolving threat of antimicrobial resistance: options for action. World Health Organization: Geneva, Switzerland, 2012. Available at: http://apps.

who.int/iris/bitstream/10665/44812/1/9789241503181_eng.pdf, Accessed: 23.09.2017 35. Allerberger, F., Gareis, R., Jindrák, V., Struelens, M. J. Antibiotic stewardship implementation in the EU: the way forward. Expert Rev. Anti Infect. Ther. 1175…1183 (2009). 36. Allerberger, F., Lechner, A., Wechsler-Fördös, A., Gareis, R. Optimization of Antibiotic Use in Hospitals … Antimicrobial Stewardship and the EU Project ABS International. 260…267 (2008). 37. Barlam, T. F. et al. Implementing an Antibiotic Stewardship Program: Guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin. Infect. Dis.62, 51-77 (2016). 38. Waterer, G. W., Wunderink, R. G. Increasing threat of Gram-negative bacteria. 75-81 (2001). 39. Zowawi, H. M. et al. The emerging threat of multidrug-resistant Gram-negative bacteria in urology. Nat. Rev. Urol. 570…584 (2015). 40. Boucher, H. W. et al. Bad bugs, no drugs: no ESKAPE! An update from the Infectious Diseases Society of America. 1…12 (2009). 41. Peleg, A. Y., Hooper, D. C. Hospital-acquired infections due to gram-negative bacteria. N. Engl. 1804…1813 (2010). 42. Falagas, M. E., Kasiakou, S. K. Toxicity of polymyxins: a systematic review of the evidence from R27 (2006). 43. Silhavy, T. J., Kahne, D., Walker, S. The Bacterial Cell Envelope. Cold Spring Harb. Perspect. 2, a000414 (2010). 44. Brown, L., Wolf, J. M., Prados-Rosales, R., Casadevall, A. Through the wal

l: extracellular vesicles in Gram-positive bacteria, mycobacteria and fungi. 620…630 (2015). 45. Morein, S., Andersson, A., Rilfors, L., Lindblom, G. Wild-type cells regulate the membrane lipid composition in a window between gel and non-lamellar structures. J. Biol. 6801…6809 (1996). 66 46. Whitfield, C., Trent, M. S. Biosynthesis and export of bacterial lipopolysaccharides. Annu. Rev. 99…128 (2014). 47. Raetz, C. R. H., Whitfield, C. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 635…700 (2002). 48. Clifton, L. A. Asymmetric phospholipid: lipopolysaccharide bilayers; a Gram-negative bacterial outer membrane mimic. J. R. Soc. Interface10, (2013). 49. Medzhitov, R., Janeway, C. J. Innate Immunity. N. Engl. J. Med. 338…344 (2000). 50. Raetz, C. R., Reynolds, C. M., Trent, M. S., Bishop, R. E. Lipid A modification systems in gram-Ann. Rev. Biochem 295…329 (2007). 51. Wang, L., Wang, Q., Reeves, P. R. The variation of O-antigens in gram-negative bacteria. 123…152 (2010). 52. Koebnik, R., Locher, K. P., Van Gelder, P. Structure and function of bacterial outer membrane proteins: barrels in a nutshell. Mol. Microbiol. 239…253 (2000). 53. Nikaido, H. Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 593…656 (2003). 54. Pagès, J. M., James, C. E., Winterhalter, M. The porin and the permeating antibiotic: a selective diffusion barrier in Gram-negative bacteria. Nat. Rev. Microbiol. 893…903 (2

008). 55. Grabowicz, M., Silhavy, T. J. Redefining the essential trafficking pathway for outer membrane 4769…4774 (2017). 56. Hirota, Y., Suzuki, H., Nishimura, Y., Yasuda, S. On the process of cellular division in : a mutant of E. coli lacking a murein-lipoprotein. Proc. Natl. Acad. Sci. 1417…1420 (1977). 57. Suzuki, H. Murein-lipoprotein of : a protein involved in the stabilization of bacterial cell envelope. Mol. Gen. Genet. 1…9 (1978). 58. Merdanovic, M., Clausen, T., Kaiser, M., Huber, R., Ehrmann, M. Protein quality control in the bacterial periplasm. 149…168 (2011). 59. Schleifer, K. H., Kandler, O. Peptidoglycan types of bacterial cell walls and their taxonomic implications. Bacteriol. Rev. 407…477 (1972). 60. Yamagami, A., Yoshioka, T., Kanemasa, Y. Differences in Phospholipid Composition between Wild Strains and Streptomycin Resistant Mutants of Certain Enteric Bacteria. Jpn. J. Microbiol. 174…176 (1970). 61. Yasuhiro, K., Yuzuru, A., Shoshichi, N. Composition and turnover of the phospholipids in 382…390 (1967). 62. Facey, S. J., Kuhn, A. Biogenesis of bacterial inner-membrane proteins. 2343…2362 (2010). 63. Okuda, S., Tokuda, H. Lipoprotein Sorting in Bacteria. Annu. Rev. Microbiol.65, 239…259 (2011). 64. Delcour, A. H. Outer membrane permeability and antibiotic resistance. Biochim. Biophys. Acta 808…816 (2009). 65. Delcour, A. H. Solute uptake through general porins. Front. Biosci. 1055…1071 (2003). 66. Masi, M., Pagès, J.

M. Structure, Function and Regulation of Outer Membrane Proteins Involved in Drug Transport in Enterobactericeae: the OmpF/C … TolC Case. Open Microbiol. J. 22…33 (2013). 67. Li, X. Z., Plésiat, P., Nikaido, H. The challenge of efflux-mediated antibiotic resistance in Gram- 337…418 (2015). 68. Richter, M. F. Predictive compound accumulation rules yield a broad-spectrum antibiotic. 299…304 (2017). 69. Hasdemir, U. O., Chevalier, J., Nordmann, P., Pagès, J. M. Detection and Prevalence of Active Drug Efflux Mechanism in Various Multidrug-Resistant Klebsiella pneumoniae Strains from Turkey. 2701…2706 (2004). 70. Vaara, M., Plachy, W. Z., Nikaido, H. Partitioning of hydrophobic probes into lipopolysaccharide bilayers. Biochim. Biophys. Acta 152…158 (1990). 71. Plésiat, P., Nikaido, H. Outer membranes of gram-negative bacteria are permeable to steroid Mol. Microbiol. 1323…1333 (1992). 72. Krishnamoorthy, G. et al. Breaking the Permeability Barrier of Escherichia coli by Controlled Hyperporination of the Outer Membrane. Antimicrob. Agents Chemother. 7372…7381 (2016). 73. Hancock, R. E. W. Alterations in Outer Membrane Permeability. Annu. Rev. Microbiol. 237…264 (1984). 74. Vaara, M. Agents that increase the permeability of the outer membrane. Microbiol. Rev. 395…411 (1992). 75. Liu, A. et al. Antibiotic sensitivity profiles determined with an Escherichia coli gene knockout collection: generating an antibiotic bar code. Antimicrob. Agents Chemot

her. 1393…1403 (2010). 76. Hancock, R. E., Bell, A. Antibiotic uptake into gram-negative bacteria. Eur. J. Clin. Microbiol. 713…720 (1988). 77. Zgurskaya, H. I., Löpez, C. A., Gnanakaran, S. Permeability Barrier of Gram-Negative Cell Envelopes and Approaches To Bypass It. ACS Infect. Dis. 512…522 (2015). 78. Silver, L. L. A Gestalt approach to Gram-negative entry. Bioorg. Med. Chem. 6379…6389 (2016). 79. Yang, N. J., Hinner, M. J. Getting Across the Cell Membrane: An Overview for Small Molecules, 29…53 (2015). 80. Wargel, R. J., Hadur, C. A., Neuhaus, F. C. Mechanism of D-cycloserine action: transport mutants for D-alanine, D-cycloserine, and glycine. J. Bacteriol. 1028…1035 (1971). 81. Chopra, I., Ball, P. Transport of antibiotics into bacteria. Adv. Microb. Physiol. 183…240 (1982). 82. Taber, H. W., Mueller, J. P., Miller, P. F., Arrow, A. S. Bacterial uptake of aminoglycoside 439…457 (1987). 83. Chopra, I. Molecular mechanisms involved in the transport of antibiotics into bacteria. 96 Suppl, 25-44 (1988). 84. Piddock, L. J. V. Clinically relevant chromosomally encoded multidrug resistance efflux pumps in bacteria. Clin. Microbiol. Rev.19, 382…402 (2006). 85. Piddock, L. J. V. Multidrug-resistance efflux pumps not just for resistance. Nat. Rev. 629…636 (2006). 86. Sun, J., Deng, Z., Yan, A. Bacterial multidrug efflux pumps: Mechanisms, physiology and pharmacological exploitations. Biochem. Biophys. Res. Commun. 254…267 (2014). 87.

Du, D., van Veen, H. W., Murakami, S., Pos, K. M., Luisi, B. F. Structure, mechanism and cooperation of bacterial multidrug transporters. Curr. Opin. Struct. Biol.33, 76…91 (2015). 88. Nikaido, H., Pagès, J. M. Broad Specificity Efflux pumps and Their Role in Multidrug Resistance of Gram Negative Bacteria. FEMS Microbiol. Rev. 340…363 (2012). 89. Okusu, H., Ma, D., Nikaido, H. AcrAB efflux pump plays a major role in the antibiotic resistance phenotype of multiple-antibiotic-resistance (Mar) mutants. J. Bacteriol.306…308 (1996). 90. Nikaido, H. Multiple antibiotic resistance and efflux. Curr. Opin. Microbiol. 516…523 (1998). 91. Rosenberg, E. Y., Ma, D., Nikaido, H. AcrD of is an aminoglycoside efflux pump. 1754…1756 (2000). 92. Zgurskaya, H. I., Nikaido, H. Multidrug resistance mechanisms: drug efflux across two membranes. 219…225 (2000). 93. Nishino, K., Yamaguchi, A. Analysis of a complete library of putative drug transporter genes in J. Bacteriol. 5803…5812 (2001). 94. Mazzariol, A., Tokue, Y., Kanegawa, T. M., Cornaglia, G., Nikaido, H. High-level fluoroquinolone-resistant clinical isolates of overproduce multidrug efflux 3441…3443 (2000). 95. Oethinger, M., Kern, W. V., Jellen-Ritter, A. S., McMurry, L. M., Levy, S. B. Ineffectiveness of topoisomerase mutations in mediating clinically significant fluoroquinolone resistance in in the absence of the AcrAB efflux pump. Antimicrob. Agents Chemother.10…13 (2000). 68 96. Webber, M. A.,

Piddock, L. J. V. Absence of Mutations in marRAB or soxRS in acrB-Overexpressing Fluoroquinolone-Resistant Clinical and Veterinary Isolates of 1550…1552 (2001). 97. Stoitsova, S. O., Braun, Y., Ullrich, M. S., Weingart, H. Characterization of the RND-Type Multidrug Efflux Pump MexAB-OprM of the Plant Pathogen Pseudomonas syringaeEnviron. Microbiol. 3387…3393 (2008). 98. Dreier, J., Ruggerone, P. Interaction of antibacterial compounds with RND eux pumps in Pseudomonas aeruginosa (2015). 99. Ziha-Zarifi, I., Llanes, C., Köhler, T., Pechere, J. C., Plesiat, P. In Vivo Emergence of Multidrug-Pseudomonas aeruginosa Overexpressing the Active Efflux System MexA-MexB-OprM. 287…291 (1999). 100. Bernal, P., Molina-Santiago, C., Daddaoua, A., Llamas, M. A. Antibiotic adjuvants: identification Microb. Biotechnol. 445…449 (2013). 101. Wright, G. D. Antibiotic Adjuvants: Rescuing Antibiotics from Resistance. Trends Microbiol.862…871 (2016). 102. Melander, R. J., Melander, C. Antibiotic adjuvants. In: Topics in Medicinal Chemistry. Springer, 103. White, A. R. et al. Augmentin (amoxicillin/clavulanate) in the treatment of community-acquired respiratory tract infection: a review of the continuing development of an innovative antimicrobial J. Antimicrob. Chemother. i3…20 (2004). 104. Leflon-Guibout, V., Speldooren, V., Heym, B., Nicolas-Chanoine, M. H. Epidemiological Survey of Amoxicillin-Clavulanate Resistance and Corresponding Molecular Mechanisms in

Escherichia Isolates in France: New Genetic Features of blaTEM Genes. Antimicrob. Agents Chemother. 2709…2714 (2000). 105. Butler, M. S., Blaskovich, M. A., Cooper, M. A. Antibiotics in the clinical pipeline at the end of 2015. 3…24 (2017). 106. Venter, H., Mowla, R., Ohene-Agyei, T., Ma, S. RND-type drug eux pumps from Gram-negative bacteria: molecular mechanism and inhibition. (2015). 107. Pagès, J. M., Masi, M., Barbe, J. Inhibitors of efflux pumps in Gram-negative bacteria. Trends 382…389 (2005). 108. Lomovskaya, O. et al. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. 105…116 (2001). 109. Pagès, J. M. Efflux Pump, the Masked Side of -Lactam Resistance in pneumoniae e4817 (2009). 110. Lomovskaya, O. et al. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. 105…116 (2001). 111. Ofek, I. Antibacterial synergism of polymyxin B nonapeptide and hydrophobic antibiotics in experimental gram-negative infections in mice. Antimicrob. Agents Chemother. 374…377 (1994). 112. Pagès, J. M., Peslier, S., Keating, T. A., Lavigne, J. P., Nichols, W. W. Role of the Outer Membrane and Porins in Susceptibility of -Lactamase-Producing Enterobacteriaceae to Ceftazidime-Avibactam. Antimicrob. Agents Chemother. 1349…1359 (2015).

113. Zabawa, T. P., Pucci, M. J., Parr, T. R., Lister, T. Treatment of Gram-negative bacterial infections by potentiation of antibiotics. 7…12 (2016). 114. Ejim, L. Combinations of antibiotics and nonantibiotic drugs enhance antimicrobial efficacy. Nat. Chem. Biol. 348…350 (2011). 115. Taylor, P. L., Rossi, L., De Pascale, G., Wright, G. D. A forward chemical screen identifies antibiotic adjuvants in ACS Chem. Biol. 1547…1555 (2012). 116. Gomez, M. J., Neyfakh, A. A. Genes Involved in Intrinsic Antibiotic Resistance of Acinetobacter 3562…3567 (2006). 117. Lee, S. Targeting a bacterial stress response to enhance antibiotic action. Proc. Natl. Acad. 14570…14575 (2009). 118. Breidenstein, E. B. M., Khaira, B. K., Wiegand, I., Overhage, J., Hancock, R. E. W. Complex Ciprofloxacin Resistome Revealed by Screening a Pseudomonas aeruginosa Mutant Library for Altered Susceptibility. 4486…4491 (2008). 119. Fajardo, A. The Neglected Intrinsic Resistome of Bacterial Pathogens. e1619 (2008). 120. Tamae, C. et al. Determination of antibiotic hypersensitivity among 4,000 single-gene-knockout mutants of J. Bacteriol. 5981…5988 (2008). 121. Lazar, S. W., Kolter, R. SurA assists the folding of outer membrane proteins. J. 1770…1773 (1996). 122. Rouvière, P. E., Gross, C. A. SurA, a periplasmic protein with peptidyl-prolyl isomerase activity, participates in the assembly of outer membrane porins. 3170…3182 (1996). 123. Hagan, C. L., Silhavy, T. J.,

Kahne, D. -Barrel membrane protein assembly by the Bam complex. 189…210 (2011). 124. Ruiz, N., Falcone, B., Kahne, D., Silhavy, T. J. Chemical conditionality: a genetic strategy to probe organelle assembly. 307…317 (2005). 125. Sklar, J. G. Lipoprotein SmpA is a component of the YaeT complex that assembles outer membrane proteins in 6400…6405 (2007). 126. Vorachek-Warren, M. K., Ramirez, S., Cotter, R. J., Raetz, C. R. A triple mutant of Escherichia lacking secondary acyl chains on lipid A. J. Biol. Chem. 14194…14205 (2002). 127. Kadrmas, J. L., Raetz, C. R. Enzymatic synthesis of lipopolysaccharide in of heptosyltransferase i. J. Biol. Chem. 2799…2807 (1998). 128. Mahalakshmi, S., Sunayana, M. R., SaiSree, L., Reddy, M. yciM is an essential gene required for regulation of lipopolysaccharide synthesis in Escherichia coliMol. Microbiol. 145…157 (2014). 129. Tsirigotaki, A., De Geyter, J., —otaric, N., Economou, A., Karamanou, S. Protein export through the bacterial Sec pathway. Nat. Rev. Microbiol. 21…36 (2017). 130. Sklar, J. G., Wu, T., Kahne, D., Silhavy, T. J. Defining the roles of the periplasmic chaperones SurA, Skp, and DegP in Genes Dev.21, 2473…2484 (2007). 131. Schäfer, U., Beck, K., Müller, M. Skp, a Molecular Chaperone of Gram-negative Bacteria, Is Required for the Formation of Soluble Periplasmic Intermediates of Outer Membrane Proteins. J. 24567…24574 (1999). 132. Grabowicz, M., Silhavy, T. J. Envelope Stress Responses: A

n Interconnected Safety Net. Trends 232…242 (2017). 133. Antonoaea, R., Fürst, M., Nishiyama, K. & Müller, M. The Periplasmic Chaperone PpiD Interacts with Secretory Proteins Exiting from the SecYEG Translocon. 5649…5656 (2008). 134. Matern, Y., Barion, B., Behrens-Kneip, S. PpiD is a player in the network of periplasmic BMC Microbiol. 251 (2010). 135. Götzke, H. et al. YfgM is an ancillary subunit of the SecYEG translocon in J. 19089…19097 (2014). 136. Pathania, R. Chemical genomics in identifies an inhibitor of bacterial Nat. Chem. Biol. 849…856 (2009). 137. Loh, B., Grant, C., Hancock, R. E. Use of the fluorescent probe 1-N-phenylnaphthylamine to study the interactions of aminoglycoside antibiotics with the outer membrane of aeruginosa 546…551 (1984). 138. Bean, G. J. A22 disrupts the bacterial actin cytoskeleton by directly binding and inducing a low-affinity state in MreB. 4852…4857 (2009). 139. Landström, J. et al. Small molecules containing hetero-bicyclic ring systems compete with UDP-Glc for binding to WaaG glycosyltransferase. 491…502 (2012). 140. Yethon, J. A., Vinogradov, E., Perry, M. B., Whitfield, C. Mutation of the lipopolysaccharide core glycosyltransferase encoded by destabilizes the outer membrane of Escherichia coliby interfering with core phosphorylation. 5620…5623 (2000). 141. Patel, D. S. Dynamics and Interactions of OmpF and LPS: Influence on Pore Accessibility and Ion Permeability. 930…938 (2016). 70 142

. Muheim, C. et al. Identification of a Fragment-Based Scaffold that Inhibits the Glycosyltransferase WaaG from Escherichia coli (2016). 143. Hajduk, P. J., Greer, J. A decade of fragment-based drug design: strategic advances and lessons Nat. Rev. Drug Discov. 211…219 (2007). 144. de Kloe, G. E., Bailey, D., Leurs, R., de Esch, I. J. P. Transforming fragments into candidates: small becomes big in medicinal chemistry. 630…646 (2009). 145. Cerveny, L. et al. Tetratricopeptide Repeat Motifs in the World of Bacterial Pathogens: Role in Virulence Mechanisms. Infect. Immun. 629…635 (2013). 146. Harms, N. The early interaction of the outer membrane protein PhoE with the periplasmic chaperone Skp occurs at the cytoplasmic membrane. J. Biol. Chem. 18804…18811 (2001). 147. Maurer, L. M., Yohannes, E., Bondurant, S. S., Radmacher, M., Slonczewski, J. L. pH Regulates Genes for Flagellar Motility, Catabolism, and Oxidative Stress in K12. 304…319 (2005). 148. House, B. Acid-stress-induced changes in enterohaemorrhagic Escherichia coli O157:H7 155, 2907…2918 (2009). 149. Stincone, A. A systems biology approach sheds new light on acid Nucleic Acids Res. 7512…7528 (2011). 150. Kannan, G. Rapid acid treatment of Escherichia coli: transcriptomic response and recovery. 37 (2008). 151. Barker, C. A. et al. Degradation of MAC13243 and studies of the interaction of resulting thiourea compounds with the lipoprotein targeting chaperone LolA. Bioorg. Med. Che